With lockdown measures being lifted in many countries, the focus of the coronavirus crisis is shifting towards how to prevent a second wave of infections. Several countries are developing smartphone apps to try and track new cases of the virus and then alert people who may have been in contact with a person infected by Covid-19. However, research by Telecompaper suggests the usefulness of the apps may be limited, as it will be difficult to reach a high penetration of users quickly.
To truly stop the spread of new infections, the tracing apps require an estimated penetration of around 60 percent of the population. Anything lower and alternative tracing systems will still be needed, such as calling people to alert them of potential exposure. This means employing more people and higher costs.
Telecompaper research suggests an app penetration of at most 20 percent is possible within a few months. Achieving this means users trust the security and privacy protection in the app and a massive public awareness campaign by state authorities.
Hoe does a tracing app work?
Use of such apps in the EU is expected to be voluntary. If someone tests positive for Covid-19, this is registered in the app. The app keeps a log of with whom the person has been in physical proximity, and these people are then alerted to their possible exposure. They can then be tested and go into quarantine if needed, stopping the spread of new infections.
Development of the apps has generated many questions, particularly in terms of privacy and security. Who has access to the data in the apps, how is it stored and for how long? Must we give up some of our privacy in the name of public health? The first attempt at developing an app in the Netherlands in just two weeks shows the difficulties in ensuring transparency and building public trust. Furthermore, it underlines the need for first designing clear use cases before developers start work and then conducting thorough security testing.
Most of the tracing apps rely on Bluetooth Low Energy to register when someone comes into proximity with another person. Apple and Google are also working with BLE, to adapt the Android and iOS operating systems for effective use of tracing apps. These changes are expected to help facilitate the development of local tracing apps.
When are tracing apps effective?
Various experts point to the apps being effective once they reach a penetration of 60 percent of the population. That's not to say they are useless under 60 percent, as they will still catch some infections, but the system needs a back-up if penetration is lower. Most often public health call centres are used then to alert people when they are affected and to help spread the word to their potential contacts for further testing.
Telecompaper data shows it is not easy for an app to reach a penetration of 60 percent. Three factors play a role: a) overall smartphone penetration, b) the penetration of Google and Apple OS and c) adoption of the app itself by end-users.
Using data from our Consumer Insights panel, we take a look at the Netherlands as an example. Smartphone penetration here is at around 92 percent. The figure is lower among people over 65, at 82 percent.
In terms of operating systems, we estimate that 95 percent of the market uses either Apple's iOS or Google's Android. Bluetooth Low Energy is supported by Apple since 2011 (from the iPhone 4S) and by Android since 2012 (Android 4). Telecompaper data shows that very few older iPhones are still used.
The Android data is a bit more difficult to establish, as the OS is used by many more brands and people often are not sure which version of the software they use. Below we show Google's global data. Based on the average replacement period for phones of 3.5 years, we can assume that most Android phones are capable of using Bluetooth LE.
The final important factor is the user. How quickly will users update their OS and download and activate a coronavirus app. In practice, many people do not update their OS often. This usually does not impact their apps, so the updates are often delayed. As for the apps themselves, Telecompaper's research shows that few apps have been able to achieve a penetration over 60 percent.
The table below shows the 10 most downloaded apps in the Netherlands and their penetration. It's important to remember that these top 10 apps have been in the market for years already and enjoy high name recognition.
These three factors illustrate how difficult it is for an app to reach a penetration of 60 percent in a short period. Policymakers and the public need to realise that other methods of contact tracing will still need to be used, requiring significant public health resources. Public authorities also will need a huge communications campaign to spread the word about the tracing apps, with an emphasis on building trust in the security of the systems and public confidence in the role they can play in halting infections.
Estimated realistic penetration
If we are very optimistic, perhaps 25 percent of people will download a tracing app within a few weeks of its release. Some weeks are need just to mount the accompanying PR campaign and reach a significant audience. Achieving 25 percent of available smartphones means a penetration of 22 percent of the population aged 16 and over. This also needs to factor in people without a smartphone or without a suitable Apple or Android device. That would put take-up in line with the app already in use in Singapore, which reached a penetration of 19 percent within a month of launch.The above analysis and experiences in other countries suggest policymakers need to look at how willing the public are to download and use such an app. Telecompaper data can provide more insight on this for various user groups, helping to develop the most effective strategy for introducing apps.
